Machine Learning Projects & Experiments

Research-driven builds demonstrating classification, detection, and forecasting expertise across domains.

See Projects View Recognitions

Research Projects

End-to-end ML solutions developed through academic research, labs, and independent exploration.

2025 · Applied Research

Mental Health Classification.

Built a multi-layer perceptron that classifies mental health conditions from user data and symptom reports.

  • Designed data preprocessing routines to clean, normalize, and encode sensitive responses responsibly.
  • Executed cross-validation with Optuna searches to calibrate depth, activation, and learning-rate schedules.
  • Delivered class probability dashboards that help support teams prioritize outreach.
View on GitHub
2024 · NLP & Trust

Detecting Fake News Using Machine Learning.

Applied NLP pipelines and supervised learning algorithms to detect misinformation in online articles.

  • Tokenized and vectorized articles using TF-IDF and word embeddings for richer semantic signals.
  • Compared logistic regression, SVM, and ensemble approaches to balance precision and recall.
  • Provided explainability reports highlighting influential terms for newsroom audits.
View on GitHub
2024 · Medical Imaging

Cancer Bacteria Detection Using Deep Learning.

Developed a CNN pipeline that identifies cancerous bacteria within microscopy data.

  • Curated and augmented medical imagery to mitigate imbalance and enable robust training.
  • Experimented with pretrained CNN backbones to accelerate convergence and reduce overfitting.
  • Integrated sensitivity metrics to maintain high recall for critical classes.
View on GitHub
2024 · Forecasting

Weather Prediction Using Deep Learning.

Forecasted temperature and precipitation trends using convolutional networks over historical data.

  • Pipeline automated feature extraction from multi-source meteorological datasets.
  • Implemented temporal CNN blocks to capture seasonality and extreme events.
  • Adopted early stopping and learning-rate schedulers to stabilize long-horizon predictions.
View on GitHub
2025 · Ensemble Learning

Multiple Diseases Prediction Using Machine Learning.

Stacked diverse models to forecast chronic disease likelihood from longitudinal health records.

  • Blended tree-based, probabilistic, and neural architectures with stacking and voting strategies.
  • Executed randomized search and Optuna optimization to elevate balanced accuracy.
  • Produced feature importance narratives to help clinicians interpret risk factors.
View on GitHub
2023 · Systems Programming

Library Management System in C.

Engineered a console-based library application that streamlines catalog tracking, lending, and member management.

  • Structured modular C code with header files for book, member, and transaction operations.
  • Implemented file-based persistence to store inventory and borrowing history without a database dependency.
  • Added search, issue/return workflows, and validation routines to keep records consistent.
View on GitHub

Tooling & Stack

Practices and platforms that keep experimentation reproducible, transparent, and scalable.

Data Preparation

Feature engineering, balancing strategies, and rigorous validation splits to reduce bias.

Model Development

Iterating with Scikit-Learn, TensorFlow, PyTorch, and ensemble frameworks for optimal fit.

Evaluation & Explainability

Tracking precision, recall, and interpretability to ensure trustworthy outcomes.

Deployment & Reporting

Packaging models with Docker, documenting via Git/GitHub, and translating results for stakeholders.

Key Outcomes

Insights gained from iterating across healthcare, safety, and environmental datasets.

Cross-Project Learnings

Translating complex data into actionable intelligence.

  • Documented reproducible workflows that accelerate future experimentation.
  • Strengthened model governance with Git-based versioning and metric tracking.
  • Produced stakeholder narratives that connect ML outputs to social impact goals.
  • Established a reusable toolkit for dataset exploration, modeling, and reporting.
See recognitions